Create app.py

app.py

@@ -0,0 +1,52 @@
+import gradio as gr
+import torch
+from transformers import BertTokenizer, BertModel
+from fastapi import FastAPI, HTTPException
+from pydantic import BaseModel
+
+app = FastAPI()
+
+class TextClassificationRequest(BaseModel):
+    text: str
+
+@app.post("/classify")
+async def classify_text(request: TextClassificationRequest):
+    # Load the pre-trained BERT model and tokenizer
+    model_name = "bert-base-uncased"
+    tokenizer = BertTokenizer.from_pretrained(model_name)
+    model = BertModel.from_pretrained(model_name)
+
+    # Preprocess the input text
+    inputs = tokenizer.encode_plus(
+        request.text,
+        add_special_tokens=True,
+        max_length=512,
+        return_attention_mask=True,
+        return_tensors='pt'
+    )
+
+    # Create a dictionary to store the output
+    output = {}
+
+    # Use the pre-trained BERT model to extract features from the input text
+    outputs = model(**inputs)
+
+    # Extract the features
+    features = outputs.last_hidden_state[:, 0, :]
+
+    # Store the output
+    output["features"] = features.tolist()
+
+    return output
+
+# Create a Gradio interface
+interface = gr.Interface(
+    fn=classify_text,
+    inputs="pdf",
+    outputs="text",
+    title="PDF Text Classification",
+    description="Upload a PDF file to classify its text"
+)
+
+# Launch the interface
+interface.launch()